Use of a genetic algorithm to optimize wheel profile geometry

Abstract

Wear is a very important subject for railway administrations. Therefore, it would be of interest to develop a methodology for designing wheel profile geometry in order to improve its behaviour in relation to this subject. Until now, existing approaches of this kind are based on the statistical study of wear wheel profiles, but this do not consider the characteristics of vehicles and railway tracks, and in addition, it is not an option when a vehicle is going to run, for example, over two different types of track over which mixed running has not previously happened. In the current paper, a new general methodology is presented for improving wheel profiles in relation to certain physical phenomena that arise during the running of the vehicle over the tracks. This methodology is based on the genetic algorithm (GA) technique. To show the power of this procedure, it is applied to an application case (of a tram-train in Spain), in which great improvements in the behaviour of wheel profile are achieved. In the light of this application, the importance of definition of indexes which control the evolution of the GA is shown.